A reinforcement learning approach to coordinate exploration with limited communication in continuous action games.
Abdel RodríguezPeter VrancxRicardo GrauAnn NowéPublished in: Knowl. Eng. Rev. (2016)
Keyphrases
- reinforcement learning
- continuous action
- policy search
- continuous state
- action space
- action selection
- function approximation
- model free
- partially observable markov decision processes
- dynamic programming
- learning agents
- game playing
- learning algorithm
- multi agent
- state space
- finite state
- game theory
- optimal policy
- robot navigation
- reward function
- reinforcement learning algorithms
- real valued
- hidden state
- sufficient conditions